Year: 2026
Type: preprint
Abstract: Untargeted LC–HRMS/MS workflows frequently represent the same molecular entity through multiple ion features, including adducts, multimers, dehydrated species, and signals compatible with in-source fragmentation. This ion redundancy can propagate into downstream analyses, increasing apparent molecular-network complexity and contributing to unannotated features. To address this issue, we developed two workflows integrated into Compound Discoverer. The unresolved workflow enables structured export of MS2 spectra in MGF format with matched quantitative data, metadata, and an audit trail, whereas the resolved workflow performs ionredundancy resolution and representative-feature selection before export. The approach was evaluated using two reference-standard panels, comprising phytocannabinoids and novel psychoactive substances (NPS), and applied to plant, biological, and in vitro drug-metabolism datasets. Ion-redundancy patterns varied across chemical classes and analytical contexts, encompassing alternative ion species, predefined mass-difference relationships, and MS2-child matches, identified when the m/z of a lower-mass coeluting feature matched a product ion in the MS2 spectrum of a higher-mass feature. Across applications, the resolved workflow reduced exported MGF entries by 18.2–36.1%. In molecular networking, the workflow reshaped molecular-family composition and connectivity through spectral-quality filtering, feature removal, and representative-ion reassignment, while preserving chromatographically distinct features lacking sufficient evidence of redundancy. These findings suggest that LC–HRMS/MS feature space does not necessarily reflect the underlying molecular space. By addressing ion redundancy before spectral export, these workflows provide a conservative and traceable framework that reduces analytical complexity, addresses a potential contributor to the apparent dark metabolome, and enables a more interpretable spectral representation that may better reflect underlying molecular diversity.
Source: ChemRxiv
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
Award: G9000845000
OpenAlex ID: W7225380395